WLC failover issue

limtohsoon · ‎11-11-2006

Hi,

I've deployed two WLC 4404-100 in a Mobility Group. Both running software version 4.0.179.8. L3 LWAPP is deployed. All the lightweight APs are configured to register to the first WLC (I call it WLC A) as Primary Controller. The second WLC (I call it WLC B) is the Secondary Controller for all the LAPs.

An SSID is created on both WLC; I call it SSID1. On WLC A, SSID1 is bridged to VLAN A on Subnet A. On WLC B, SSID1 is bridged to VLAN B on Subnet B.

During normal operations, client associates to SSID1. It is assigned IP address on Subnet A. When WLC A fails, all APs re-register to WLC B. Client refreshes its IP address and is assigned IP address on Subnet B. Looks fine.

When WLC A is back online (AP Fallback is enabled), all APs register back to WLC A. But client is still on Subnet B and connectivity is okay. Doing "ipconfig /release" and then "ipconfig /renew", client is still getting back Subnet B.

Can you explain why client is not bridged to Subnet A when the APs fall back to WLC A?

Thank you.

B.Rgds,

Lim TS

zhenningx · ‎11-12-2006

Although all the APs fall back to WLC A, the client's anchor is still WLC B. During the AP register back to WLC A process, the clients behavior is like roaming. So it still get the same IP on the anchor WLC (B). If you remove the client from WLC B and do "ipconfig/release" and "ipconfig/renew" again, they should get IP from subnet A.

limtohsoon · ‎11-12-2006

Hi,

Thanks for your reply.

During normal operations, all LAPs are registered to WLC A. When WLC A fails, client gets disconnected for a brief moment, and then refreshes its IP address automatically, and is now on Subnet B. In this scenario, why isn't it getting the same Subnet A IP address because it initially anchored to WLC A?

When APs fallback to WLC A, I removed the client on WLC A and it got IP from Subnet A again.

Thank you.

B.Rgds,

Lim TS

zhenningx · ‎11-13-2006

Lim, the WLC crash and AP fall back are different. When WLC A crashes, it can not send mobility update to other WLCs in the mobility group about it's local clients before it goes down. So clients can not successfully roam to other WLCs. They have to receive another IP and reauthenticate. When AP fall back between WLCs. WLCs and communicate between each other about the clients mobility information successfully, so client can keep the same IP and no need to reauthenticate. Hope my answer helps.